Serms Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

Inflammation ( # 2018) DOI: 10.1007/s10753-018-0763-1

ORIGINAL ARTICLE

SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

Lauri Polari ,1,4 Anu Wiklund,1 Sofia Sousa,1,2 Lauri Kangas,3 Tero Linnanen,3 Pirkko Härkönen,1 and Jorma Määttä1

Abstract— Signaling via estrogen receptors (ER) is recognized as an essential part of the immune regulation, and ER-mediated signaling is involved in autoimmune reactions. Espe- cially ERα activation in immune cells has been suggested to skew cytokine production toward Th2/M2-type mediators, which can have protective effect on inflammatory diseases and reduce Th1 and Th17 responses. These effects are caused by increased alternative activation of macrophages and changes in the activation of different T cell populations. In humans, hormonal status has been shown to have a major impact on several inflammatory diseases. Selective estrogen receptor modulators (SERMs) are ER ligands that regulate ER actions in a tissue-specific manner mostly lacking the adverse effects of steroid hormones. The impact of SERMs on the immune system is less studied, but it is suggested that certain SERMs may also produce immunoprotective effects. Here, we show that two novel SERMs and raloxifene affect immune cells by promoting M2 macrophage phenotype, alleviating NFκB activity, inhibiting T cell proliferation, and stimulating the production of anti- inflammatory compounds such as IL10 and IL1 receptor antagonist. Thus, these compounds have high potency as drug candidates against autoimmune diseases.

KEY WORDS: estrogen receptor; SERM; macrophages; inflammation; raloxifene; estradiol.

INTRODUCTION source of immune stimulus, and variability of expression of ER subtypes in the cellular microenvironment [7]. Estro- Signaling via estrogen receptors (ER) is recognized as genic compounds such as female sex hormones elicit their an essential part of the immune regulation, and ER- effects via ER. Upon ligand binding, ER initiates gene mediated signaling involved in both chronic inflammatory transcription in the nuclei and also elicits immediate effects diseases and autoimmune reactions [1–6]. This regulation via cytosolic signaling cascades. ER have been utilized as a can be either pro- or anti-inflammatory depending on sev- drug target for several estrogen-regulated diseases, most eral criteria such as types of organs and cells involved, importantly breast cancer and osteoporosis, in estrogen- sensitive organs [8]. However, ER-modulated inflammatory diseases and autoimmune reactions are not only limited Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10753-018-0763-1) contains supplementary ma- to traditional estrogen target tissues. terial, which is available to authorized users. Estrogenic signaling is suggested to affect immunomodulation in a wide array of inflammatory dis- 1 Institute of Biomedicine, University of Turku, Turku, Finland 2 Faculté de Médecine, Université Lyon-1, Lyon, France eases such as intestinal inflammation and CNS autoimmu- 3 Forendo Pharma Ltd., Turku, Finland nity [7]. ER ligands possibly induce anti-inflammatory 4 To whom correspondence should be addressed at Institute of Biomedi- effects via mechanisms involving ERα and GPER cine, University of Turku, Turku, Finland. E-mail: [email protected]

0360-3997/18/0000-0001/0 # 2018 The Author(s) Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä activation on immune cells, inducing a Th2-type skew in patterns of cytokine production and receptor expression, the cytokine milieu and reducing Th1 and Th17 responses according to their microenvironmental cues [24]. Several [1, 9–13]. This anti-inflammatory shift includes increased studies suggest that E2 activation modulates macrophage M2 characteristics in monocyte macrophage populations phenotype and function [12, 16, 25–28] although E2 does and changes in the activity and number of T regulatory not elicit strong proliferative response in myeloid cells in a cells (Treg) [14–17]. It is intriguing that a similar Th2-type similar degree as in classical estrogen-sensitive tissues skew in inflammatory mediators has been observed to such as breast epithelium and endometrium [19]. Never- occur at the third trimester of pregnancy—aperiodalso theless, recent studies suggest that endogenous E2 may characterized by increased estrogen levels [18]. These ob- also induce macrophage activity and self-renewal in vivo servations suggest that ER signaling regulates the immune [16, 26, 29]. The expression of ER in myeloid cells has system cells by modulating their responses to inflammato- been demonstrated, but the pattern of ER subtypes and ry stimuli. splice variants in macrophages and their possible contribu- The activation of ER signaling is considered to stim- tion to the innate immunity is likely to be context/tissue ulate anti-inflammatory response. Accordingly, 17β- specific [16, 28, 30, 31]. estradiol (E2), a strong ER agonist steroid hormone, is We have studied the immune-modulatory properties associated with amelioration of inflammatory diseases of SERM drugs and several novel SERM candidates and [7]. E2 is not, however, utilized as an immunomodulatory examined their effects on NFκB activity and macrophage drug because it may increase a risk for tumor formation in surface protein expression in human monocytes. Among estrogen-sensitive tissues [19]. Therefore, it is not an opti- those compounds, we have selected the most M2-type mal drug for long-term medication. Instead, a recent clin- activation promoting compounds to be further studied in ical trial suggests that another natural estrogen hormone, models of macrophage differentiation and cytokine pro- estriol, may be more suitable as an immune modulator. duction. Here, we present evidence that specific SERM Estriol, the natural level of which is high especially during compounds can modulate the function of the monocyte pregnancy, together with glatiramer acetate, has been macrophage cell population. We show that specific SERMs shown to reduce the relapse rate of female MS patients can modulate the differentiation of CD14-positive cells by [14, 20]. This result is supported by clinical findings show- skewing their phenotype toward M2 macrophages and ing that symptoms of multiple sclerosis are often alleviated inducing the production of anti-inflammatory cytokines. during pregnancy [21]. These novel, previously unpublished SERM compounds Selective estrogen receptor modulators (SERMs) are promoted M2-like macrophage activation in pharmacolog- compounds that elicit estrogenic and/or antiestrogenic ef- ical concentration. They could thus be beneficial for the fects in a tissue-specific manner. The first non-steroidal alleviation of systemic inflammation in autoimmune SERM developed for clinical use was tamoxifen, which diseases. has been used as an endocrine therapy for breast cancer since the late 1980s [22]. Thereafter, a few other SERM drugs such as raloxifene, toremifene, and ospemifene have MATERIALS AND METHODS been developed for treatment and/or prevention of various diseases in estrogen-responsive tissues. As the activation of SERMs ERsinimmunecellshasbeenlinkedwithdownregulation All SERM compounds, including novel compounds of inflammation, it is tempting to expect that an appropriate investigated in this study, were a kind gift from Forendo SERM might regulate the ER signaling in immune cells but Pharma Ltd., Turku, Finland, unless stated otherwise. Ral- not induce adverse effects in other estrogen-sensitive oxifene and 17β-estradiol were purchased from Sigma (St. tissues. Louis, USA). Working dilutions of estrogenic compounds Monocytes and macrophages are myeloid antigen- were prepared in dimethylsulfoxide and kept at − 20 °C. presenting cells that have a central role in the initiation of immune responses thus possessing a major role in adaptive Estrogen Receptor Affinity Assay immunity. In addition to phagocytosis, they can either promote or alleviate immune responses [23]. Currently, it The binding affinity of SERM2, SERM7, and ralox- is widely accepted that rather than just pro-inflammatory ifene to ERα and their ability to compete in binding with M1 and anti-inflammatory M2 dichotomy, the macro- E2 (incubation time 2 h, estrogen concentration 2 nM) phages can adapt a spectrum of phenotypes with different were determined using a commercially available assay kit SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells according to the manufacturer’s instructions (PanVera (IL4) (50 ng/ml), or IL10 (50 ng/ml) [36]. M(IFNγ)cells LCC, Madison, WI) [32]. The IC50 values were obtained were activated with lipopolysaccharide (LPS, 10 ng/ml) by fitting the data to the Hill’sequation[33]. after 5-day incubation with IFNγ. Monocytes were treated with SERMs or E2 to assess Cell Culture their effects on macrophage polarization and phenotype. Compounds were added 4 days after the differentiation Human THP-1 Lucia NFκB reporter cells were pur- stimuli (IFNγ), followed by LPS addition 24 h later. After chased from InvivoGen (San Diego, CA, USA). The THP- 6 days of culturing, medium samples were collected and 1LuciaNFκB cell line contains a stable NFκB inducible cells were either lyzed for RNA extraction with RA1 lysis Lucia reporter construct which is secreted to the growth buffer (Macherey-Nagel, Duren, Germany) supplemented medium. THP-1 Lucia cells were grown in a colorless with β-mercaptoethanol or detached (with Accutase) for RPMI-1640 growth medium with 4.5 g/l glucose, 10 mM flow cytometry analysis. Cell mediums and lysates were HEPES, 1.0 mM sodium pyruvate, 2 mM L-glutamine, kept at −80 °C prior to RNA or protein quantitation assays. Pen-Strep (50 U/ml, 50 mg/l), 100 mg/l Normocin (InvivoGen), and 10% heat-inactivated fetal bovine serum (iFBS, EU approved). All reagents unless stated otherwise Subjects were purchased from Thermo Fisher Scientific Life Tech- All subjects donating blood were volunteers, in- nologies (Carlsbad, CA, USA). Cells were maintained by formed about the study, and they gave a written consent 6 passing them every 3–4 days by inoculating 0.5–1×10 on the use of their cells. No personal information of the cells to fresh medium. To maintain selection pressure, subjects was recorded. We used blood cells from total six every other passage was supplemented with 100 mg/l healthy, non-obese males with age range 20–50 years. The Zeocin (InvivoGen). All reagents used in cell culturing experiments with human cells were done under approval of were purchased from Thermo Fisher Scientific Life Tech- the University of Turku Ethics Committee (Statement Ref. nologies (Carlsbad, CA, USA) unless stated otherwise. 6/2017), and studies were conducted in accordance with MCF-7 and Ishikawa estrogen responsive element the Helsinki Declaration. (ERE) Luc reporter cells were transfected with the reporter construct ERE Luc containing two tandem consensus ERE Cell Viability Assay cloned into the pGL2-Promoter cloning vector. Cells were grown and maintained in DMEM medium supplemented The cytotoxicity of SERMs on CD14+ and THP-1 with 10% iFBS, 2 mM L-glutamine, 150 μM G-418 for Lucia cells was evaluated with the AlamarBlue cell viabil- selective pressure, 10 μg/ml insulin, and 1 nM E2 (two ity reagent (Thermo Fischer) according to the manufac- latter ones only for MCF-7). Other methods were similar to turers’ instructions. Briefly, cell viability was determined Barsalou et al. [34] and Kallio et al. [35]. by culturing cells at least in triplicate in 100 μlofmedium (5 × 105 cells/well) for 2 days in flat-bottomed 96-well Primary Cell Isolation and Differentiation plates with respective treatment. After incubation, 10 μl of AlamarBlue reagent was added to each well and the Peripheral venous blood was obtained from healthy microtiter plate further incubated for 2 h at 37 °Cin5% males, and leukocytes were separated using a Ficoll-Paque CO2. Then, the fluorescence was read with a HIDEX Plate (GE Healthcare Life Sciences, Uppsala, Sweden) solution. Chameleon Reader (Hidex Ltd., Turku, Finland). Peripheral blood mononuclear cells (PBMC) were collected and washed with PBS after centrifugation. Monocytes Luciferase Reporter Activity Assays were isolated by CD14-positive selection using anti-CD14 magnetic beads (Miltenyi Biotech, Lund, Sweden). Mono- THP-1 Lucia reporter cells were grown on 96-well cytes were suspended in colorless αMEM supplemented plate (1 × 105/well) 24 h in the presence of SERM followed with glutamine, penicillin–streptomycin, and 10% by 24 h with LPS or tumor necrosis factor alpha (TNFα, iFBS (EU). Monocytes were plated at 5 × 105 cells Life Technologies) stimulation. Cell culture medium per well in clear 12-well plates. Monocytes were samples were collected for reporter activity assay, and cultured in fresh medium for 6 days at 37 °C to allow cell viability was measured as described in the previous differentiation into macrophages. Resting cells were polar- chapter. Cells were washed and freeze-thaw lyzed for total ized into M1 or M2 type of macrophages by incubation for protein assay. Luciferase activity was determined by 6 days with interferon-γ (IFNγ, 50 ng/ml), interleukin-4 relocating 20 μl aliquots of cell culture media into optical Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä

Lumitrac 96-well plates (Greiner Bio-One, Krensmünster, antibodies (from BioLegend, San Diego, CA, USA, unless Austria) followed by Quanti-Luc luciferase substrate stated otherwise: CD14 APC/Cy7 (clone63D3), CD163 (InvivoGen). Plates were read immediately for luciferase Alexa 647 (clone GHI/61; BD Biosciences), CD192 activity with Victor2 multiplate reader (PerkinElmer). BV605 (clone K036C2), and CD206 Alexa Fluor 488 To study ERE Luc activation, MCF-7 and Ishikawa (clone 15-2)). Cells were washed with PBS prior to anal- reporter cells were pre-incubated in colorless DMEM me- ysis, and the expression of cell surface proteins was ana- dium supplemented with charcoal-stripped serum to re- lyzed. Unstained macrophages were included as negative move endogenous steroids without adding insulin or E2 control for staining. Data were analyzed with Flowing [35]. Cells were plated at densities of 400,000 (Ishikawa) Software (Turku Centre of Biotechnology, Turku, Fin- or 200,000 (MCF-7) per 96-plate well. After 24-h incuba- land), after gating on the myeloid population in the FSC/ tion the media were replaced with fresh ones with or SSC plot with 10,000 events recorded. Results were without the study drugs and/or 1 nM of E2 dissolved in expressed as the relative values of the median fluorescence DMSO. After 48 h, media were again removed and re- intensity (MedFI) of each fluorochrome normalized to the placed with DMEM. Fifty microliters of freshly respective value of unstained control. The percentage of reconstituted luciferase substrate was added into each well. cells expressing a specific phenotype was calculated ac- After a 5-min incubation at room temperature, the plate cording to the number of cells in the dot plot quadrants was measured with Victor multilabel counter. after gating.

RNA Extraction and Gene Expression Analysis Macrophage Secretion Analysis RNA was extracted from lyzed cell samples by using Concentrations of inflammatory regulators IL1β, the Nucleospin RNA kit (Macherey-Nagel) according to IL10, chemokine (C-C motif) ligand 2 (CCL2), interleukin the manufacturers’ instructions. Total RNA concentration 1 receptor antagonist (IL1RA), and TNFα in cell culture in samples was determined by using Nanodrop ND-1000 medium were analyzed with multiplex assays (eBioscience (Nanodrop, Wilmington DE, USA). RNA sample was Ltd., Cheshire, UK) according to manufacturer’sinstruc- eluted in RNAse free water and translated to cDNA using tions by using the Luminex 200 system (Luminex Corpo- Sensifast cDNA Synthesis Kit (Bioline, London, UK). ration, Austin, TX). Medium samples were collected after Quantitative PCR was performed on a CFX96 thermal 6 days of culturing CD14-positive cells (500,000 cells in cycler (Bio-Rad Laboratories, Hercules, USA) with 1 ml of medium per sample) and kept at − 80 °Cuntil Taqman Gene Expression Assays (Applied Biosystems, analyzed. To remove debris, samples were centrifuged Foster City, CA, USA). Eight nanograms of total RNA (5 min, 7000 rpm) prior to the assay. For IL1RA and was used per sample unless stated otherwise, and the assay CCL2 assays, samples were first diluted 1:100 in culture was performed according to the manufacturers’ instruc- medium. tions. The delta-delta Ct method was used for relative quantification of gene expression. Glyceraldehyde 3- T Cell Stimulation and Proliferation Assays phosphate dehydrogenase (GAPDH) was used as a refer- PBMCs were isolated from heparinized blood as de- ence gene. Taqman primer information is provided in scribed in the previous chapter. The cell suspension was supplementary Table 1. diluted to a final concentration of 106 cells/ml and plated in flat-bottomed 96-well culture plates, 1 × 105 cells/well. Flow Cytometry Analysis Cells were activated with 10 μg/ml phytohemagglutinin Macrophages and monocytes were analyzed with the (PHA-P) from Phaseolus vulgaris (Sigma-Aldrich, St. BD LSR Fortessa flow cytometer (BD Biosciences, Louis, MO, USA). Plates were incubated for 3 days at + Erembodegem-Dorp, Belgium). Cells were detached by 37 °Cin5%CO2. Non-adherent cell proliferation was Accutase (Life Technologies), washed with PBS with measured by the thymidine incorporation method. After 0.1% FBS, and filtered through 35-μm pore cell-strainer incubation, 1 μCi [6-3H]-thymidine (PerkinElmer, Boston, snap caps (Corning Incorporation, Durham, NC, USA). MA, USA) was added to each well, followed by 20 h of Before staining, cells were transferred to a 96-well plate incubation. Cell suspensions were directly transferred to and Fc receptors were blocked with Human BD Fc Block MultiScreen HTS FB 96-well filter plate (Merck Millipore, reagent (BD Biosciences). Cells were then incubated 1 h Darmstadt, Germany), and the wells were washed twice with the fluorochrome-labeled anti-human monoclonal with Milli-Q-grade water and dried thoroughly. Optiphase SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

HiSafe 3 (PerkinElmer, Waltham, MA) scintillation cock- antagonism in both reporter cell lines, but no SERM7- tail was added in the wells, and the plate was sealed. After induced agonism was detected in either of them. 30-min incubation at RT, the plate was counted on a Wallac 1450 Microbeta Trilux β-counter (PerkinElmer, Turku, Selective Estrogen Receptor Modulators Suppress Finland) for 1 min per well. NFκB Activation To elucidate the role of ERs on NFκB activation in monocytes, THP-1 Lucia cells with stable NFκBreporter RESULTS expression were grown 24 h in the presence of SERM compounds, followed by 24 h with LPS or TNFα stimu- Characterization of Two Novel SERMs lation. First, several novel compounds and SERM drugs including tamoxifen, ospemifene, raloxifene, fulvestrant, We characterized the abilities of two novel com- and fispemifene were pre-screened for their NFκBmodu- pounds (referred here as SERM2 and SERM7) to activate latory activity, and the three most efficient ones were or inhibit ERE-mediated signaling in two estrogen- selected for a more specific analysis. The novel compound ’ sensitive reporter cells and the compounds ability to com- SERM7 inhibited both LPS- and TNFα-induced reporter pete for ER binding with E2 [37][32]. As shown by the activity in a concentration-dependent manner (Fig. 1a). α higher IC50 value, SERM2 bound to ER with a slightly Raloxifene and SERM2 also downregulated TNFα- lower affinity than raloxifene and E2 (Table 1). The affin- stimulated, but not LPS-stimulated, reporter activity (Fig. α ities of E2 and raloxifene to ER were of similar magni- 1b, c). In fact, raloxifene slightly increased LPS-stimulated tude to previously reported ones [38]. SERM7 binding activity. Both LPS and TNF alone produced an over 10- α affinity to both ER was considerably weaker than that fold increase in reporter activity. Basal reporter activity was of raloxifene. detectable, but none of the SERM compounds had signif- The human breast cancer cell line MCF7 and the icant effects on it (data not shown). Esr1, Esr2,andGper1 human endometrial cancer cell line Ishikawa stably ex- genes were all expressed in THP-1 Lucia cells (Fig. 2a). pressing ERE Luc reporter constructs (MCF-7 ERE Luc reporter and Ishikawa ERE Luc reporter cells, respectively) SERM Compounds Increase the M2 Phenotype were used to study ERE-mediated transcriptional activa- Characteristics of CD14-Positive Macrophages tion by the SERMs in vitro. Both of these cell lines express ERα and ERβ and are commonly used to study estrogenic To further study whether the NFκBreporterattenua- signaling [39–41]. At the concentrations higher than tion was related to macrophage phenotypic changes in the 100 nM, SERM2 was found to activate ERE in Ishikawa M1-M2 axis, we measured the effect of SERMs on the cells acting as an agonist. In the presence of E2, SERM2 macrophage surface marker expression by flow cytometry. antagonized Ishikawa ERE Luc reporter activity, at the To obtain macrophages of a specific phenotype, we cul- same order of potency as raloxifene (Table 1). An agonistic tured peripheral blood-derived CD14+ monocytes in the action of SERM2 was not found in MCF-7 ERE Luc presence of IFNγ, IL4, or IL10. The expression of the reporter cell, but SERM2 antagonized ERE Luc activity Esr1, Esr2,andGper1 was confirmed by gene expression in MCF7, although with a considerably higher IC50 value analysis, Esr1 being over 10-fold higher to others in both when compared to raloxifene. SERM7 displayed weak ER M(IFNγ) cells and non-activated primary monocytes (Fig.

Table 1. Characterization of Estrogenic Properties of Novel SERMs in Comparison to Raloxifene and 17β-Estradiol

Binding affinity to ERα MCF-7 ERE Luc activity Ishikawa ERE Luc activity

Agonism Antagonism Agonism Antagonism

SERM2 +++ 0 ++ + +++ SERM7 + 0 + 0 + Raloxifene ++++ 0 +++ 0 +++ 17β-Estradiol ++++ ++++ − ++++ −

Incubation time and estrogen concentration were 2 h and 2 nM for affinity assays and 48 h and 1 nM for reporter assays, respectively ++++ IC50 0.5–2nM,+++ IC50 < 10 nM, ++ IC50 < 100 nM, + IC50 < 1 μM, 0 no estrogenic effects Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä

Fig. 1. SERMs modulate NFκB activity in monocytic THP-1 Lucia reporter cells. Effects of a SERM7, b SERM2, and c raloxifene on 24-h LPS or TNF- induced reporter activity of THP-1 monocytes. c(LPS) = 10 ng/ml, c(TNFα) = 5 ng/ml. Dots and error bars represent mean ±SD. Statistical significance was determined after unpaired t test vs control (0 μM). *P <0.05;**P < 0.01; ***P < 0.001.

2b, c). Only a small percentage of IFNγ-induced macro- cell number, but 17β-estradiol (E2) at 10 nM concentration phages express the surface proteins CD206 (mannose re- caused (statistically not significant) stimulation but to a ceptor, MRC-1) and CD163 compared to IL4- and IL10- lesser extent than SERM2. stimulated monocytes (supplementary Figs. 1–2), and both Raloxifene and SERM7 both increased the gene ex- CD163 and CD206 are considered strong M2 markers [17, pression of CD206 but not of CD163 (Fig. 3b, c). Never- 42]. Addition of 1 μM SERM2 to the culture, 24 h prior to theless, raloxifene increased surface protein expression of LPS activation, increased the relative proportion of CD14+ scavenger receptor CD163 as demonstrated by its in- CD163+ CD206+ macrophages (fold change 1.7; P = creased median fluorescence intensity (MFI) but reduced 0.025, Fig. 3a). This was supported by a significant eleva- CD206 MFI (fold changes 1.4; P <0.001 and 0.7; tion of CD206 mRNA expression (fold change 2.2; P < 0.001, respectively, Fig. 4a, b). Under the same point P < 0.001, Fig. 3b) and an observed increase in CD163 of observation, the level of CD206 mRNA was, however, mRNA expression although not statistically significant increased (fold change 2.0; P = 0.025, Fig. 3b). We also (fold change 1.2; P =0.12,Fig.3c). Other studied SERMs studied the expression of a few other macrophage subtype- did not have an effect on relative CD14+ CD163+ CD206+ associated surface receptors. CD192 (CCR2) is considered

Fig. 2. Expression of estrogen receptor subtypes (Esr1 and Esr2)andGper1 genes in a THP-1 Lucia cells, b CD14-positive primary M(IFNγ+LPS) cells, and c non-activated monocytes. Boxes extend from min to max, and line in the middle represents median. SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

Fig. 3. SERMs promote an alternative macrophage activation of monocytes by inducing activation of anti-inflammatory CD14+ CD163+ CD206 macrophage phenotype. a Proportion of CD14+ CD163+ CD206+ macrophages from CD14+ cell population. b Gene expression of CD206 and c CD163 in CD14+ M(IFNγ+LPS) cells. Box extends from the 25th to 75th percentiles; line in the middle represents median, and whisker from min to max. One-way statistical significance was determined after one-way ANOVA and Holm-Sidak post hoc multiple comparison test. *P <0.05;**P <0.01; ***P <0.001. d, e Representative FACS dot plots (CD163-AF647 vs CD206-AF488) showing M1 (lower left quadrant) and double-positive CD163+ CD206+ (upper right quadrant) M2-like cells after IFNγ+LPS.

to be an M2 marker, and both SERM2 and SERM7 in- change 0.6 and 0.8, respectively, P < 0.001 for both, Fig. creased moderately its MFI (Fig. 4c) whereas they signif- 4d) that is a component of the membrane protein complex icantly inhibited expression of CD14 surface protein (fold responsible for pathogen detection and LPS-triggered pro- Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä

Fig. 4. Effect of 48-h SERM/E2 treatment on median fluorescence intensity representing surface receptor expression of a CD206, b CD163, c CD192, and d CD14 in human-derived CD14-positive mononuclear cells, cultured 6 days for monocyte polarization and IFNγ 50 ng/ml followed by LPS activation at day 5. Scatter plot of individual samples with means ± SD. Statistical significance vs control group was determined after one-way ANOVA followed by post hoc Holm-Sidak multiple comparison test. *P <0.05;**P < 0.01; ***P < 0.001. inflammatory responses [43]. E2 treatment also had a 1.4; P = 0.05) treatments. Mean IL1RA protein secretion reducingtrendonmeanCD14expression(Fig.4d) rate was increased, respectively, but these results failed to (Table 2). reach the significance due to considerable variation (fold change 1.4; P = 0.19 for SERM2, Fig. 5d). Novel Selective Estrogen Receptor Modulators Induce the Production of Th2-Type Cytokines IL10 and SERM-Induced Macrophage Polarization Toward a IL1RA M2 Phenotype Does Not Inhibit Inflammatory Signaling After the flow cytometric studies, we wanted to assess whether the cytokine production of CD14+ M(IFNγ)cells As our results suggested that SERMs skew macro- can also be skewed toward the Th2/M2 phenotype by phage phenotype toward M2 (Figs. 3, 4,and5), we were SERM treatment. We found that 1 μM SERM2 and ralox- interested in whether expression of the genes for and ifene and 10 nM E2 increased IL10 mRNA expression in secretion of pro-inflammatory mediators is attenuated to M(IFNγ) cells (fold change 1.3 for all; P = 0.026, 0.041, same extent. Neither SERMs nor E2 affected the expres- and 0.05, respectively, Fig. 5a). Concomitantly, IL10 pro- sion of the pro-inflammatory genes IL1β, TNFα, IL6, tein secretion from cells was increased as demonstrated by IL12B, and TLR4 in LPS-activated M(IFNγ) cells (Fig. multiplex analysis from cell growth media, SERM2 having 6a–e). Respectively, secretion of IL1β and TNFα from the most significant effect on IL10 secretion (fold change CD14+ cells to growth medium was not changed (Fig. 1.4; P = 0.014, Fig. 5b). In addition to IL10, the gene 6f, g). The only gene responsive to ER ligands and con- expression of IL1RA was upregulated after SERM2 (fold sidered to be pro-inflammatory was CCL2, whose expres- change 1.5; P =0.05,Fig. 5c) and SERM7 (fold change sion was moderately downregulated by SERM7 and SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

Table 2. Summary of the Statistically Significant Effects of SERM2, SERM7, and Raloxifene on Macrophage Polarization and Activity Markers, Presented in Figs. 1, 2, 3, 4, 5, 6,and7

SERM2 SERM7 Raloxifene

NFκB activity (LPS-induced) n − + NFκB activity (TNFα-induced) −−− Number of CD163+ CD206+ double-positive cells + n n CD14 surface expression −−n CD163 surface expression n n + CD163 gene expression n n n CD192 surface expression + + n CD206 surface expression + + − MRC1 gene expression + + + IL10 expression + n + IL10 secretion + n n IL1RN expression + + n IL1RA secretion n n n CCL2 expression n −− CCL2 secretion n n n PHA-activated T cell proliferation −−−

+ significant increase, − significant decrease, n no effect

Fig. 5. Estrogen receptor ligands modulate cytokine synthesis and secretion of (IFNγ+LPS)-activated CD14+ cells toward anti-inflammatory phenotype. a Secretion and b gene expression rates of IL10. c, d Respective values for IL1RA. Five hundred thousand cells per sample. Box extends from the 25th to 75th percentiles and whiskers from min to max. Statistical significance was determined after one-way ANOVA followed by post hoc Holm-Sidak multiple comparison test. #P ≤ 0.1; *P <0.05. Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä

Fig. 6. Estrogen receptor ligands do not downregulate proinflammatory signaling in (IFNγ+LPS)-activated CD14+ macrophages. Gene expression rates of a IL1β, b TNFα, c IL6, d IL12B, and e TLR4. Secretion rates of f TNFα and g IL1β from M(IFNγ+LPS) cells. h Gene expression and i secretion rates of CCL2. Five hundred thousand cells per sample. Box and whisker as min and max. Statistical significance was determined after one-way ANOVA followed by post hoc Holm-Sidak multiple comparison test. *P <0.05.

raloxifene (fold change 0.8 for both). Nevertheless, the SERM2, SERM7, and raloxifene all inhibited proliferation attenuation of gene expression was not translated to protein of activated T cells (fold changes 0.4, 0.7, and 0.7, respec- secretion from the same cells (Fig. 6h, i). tively, Fig. 7). SERM7 suppressed proliferation rate at lower those than other SERMs here. The mean proliferation rate of E2-treated cells was below control, but there SERMs Inhibit T Cell Proliferation Rate were substantial differences in E2 response between the To assess whether the SERM-induced macrophage cells from different donors. Interestingly, corresponding polarization affected T cell activation, we measured the donor-associated effects were not found with other treat- rate of T cell proliferation in co-cultures with PBMC. T ments or assays. SERMs (0.3–3 μM) and E2 (10 nM) alone cell proliferation was activated with the exogenous at the concentrations used did not possess cytotoxic or superantigen PHA, followed by a 48-h SERM treatment. growth stimulatory effects on T cells nor monocytes as cell SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

Fig. 7. SERMs inhibit proliferation of PHA-activated primary T cells. Peripheral blood-derived mononuclear cells from five donors were cultured 3 days with phytohemagglutinin (PHA) and SERM or E2. Proliferation rate of non-adherent cells was measured with β-counter after 20 h 3H-thymidine incorporation. Selective estrogen receptor modulator stimulates IL-10 secretion from IFNγ+LPS-activated CD14-positive monocytes. Box extends from the 25th to 75th percentiles; line in the middle represents median, and whisker from min to max. Statistical significance was determined after one-way ANOVA followed by post hoc Holm-Sidak multiple comparison test. *P <0.05;**P < 0.01; ***P < 0.001.

viabilities were not changed by treatments (data not gene expression of the M2 markers after SERM2 treat- shown). ment. We found that different SERMs activate M2 markers in to a variable degree, as raloxifene was the most potent CD163 activator whereas SERM2 and SERM7 significant- DISCUSSION ly upregulated CD206 and CD192. Second, SERM2 and raloxifene increased the expression and secretion of the Here, we present two new SERM-type compounds, anti-inflammatory mediators IL10 and IL1RA that are which promote alternative M2 activation of myeloid cells. suggested to inhibit pro-inflammatory signaling, antigen According to our results, the macrophage phenotype- presentation, and the CD4+ effector T cell activation. IL10 polarizing anti-inflammatory activity of these new com- and IL1RA are considered to be anti-inflammatory medi- pounds is higher than those of previous SERM drugs. ators whose synthesis and secretion are associated with the Macrophages and monocytes play a critical role in the M2-subtype macrophages (Supplementary Fig. 3)[42, 47]. pathogenesis of several autoimmune diseases by acting as Third, SERM-induced macrophages suppressed T cell pro- key producers of mediators of inflammation further liferation after antigen activation in primary PBMC cul- supporting the development of CD4+ and CD8+ pro- ture. Fourth, SERM2, SERM7, and raloxifene suppressed inflammatory cells [44]. The dysregulation of macro- the TNFα-induced NFκB activation of the reporter mono- phages is found in various tissue-specific pathologies such cytes. Our data demonstrates that the selected SERMs as neuronal degeneration in Alzheimer’sdiseaseandhy- contribute to M1/M2 macrophage activation by promoting perplasia in synovial joints in rheumatoid arthritis [45, 46]. anti-inflammatory properties in M1-like cells thus stimu- This study provided several lines of evidence that lating production of Th2-type compounds. SERM7, raloxifene, and most efficiently SERM2 promot- Our results are in line with the current perspective ed M2-type characteristics in IFN + LPS-activated macro- that, instead of the simplified M1/M2 dichotomy, macrophages. First, the capacity of SERM2 to increase the num- phage polarization includes numerous states of activation ber of CD14+ CD163+ CD206+ cells clearly suggests that with diverse functions [24, 48]. Interestingly, we did not a proportion of IFNγ-polarized macrophages were able to find distinct evidence that the SERMs tested would directly adopt a more anti-inflammatory phenotype even in the downregulate pro-inflammatory cytokine signaling in presence of LPS. There was a concomitant increase in the CD14+ cells. In the cells where IL10 and IL1RA were Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä upregulated, the expression and secretion of IL1β and with these partly conflicting data of the E2 having diverse, TNFα remained unaltered. To confirm this result, we also context, or even donor-specific effects on immune system measured the expression of a few additional inflammatory responses. We found that the effect of E2 on CD14+ cells genes such as IL12B, IL6, and TLR4, but no SERM- varied from modest to non-existent and E2 induction on the induced downregulation was found either. CCL-2 was an proliferation rate of activated T cells varied significantly exception among pro-inflammatory genes as its expression between the cells from different donors. Similar variation was downregulated by SERM7 and raloxifene. SERM2 was not found among SERM-stimulated cell samples. A has the most significant stimulatory effect on M2 markers, recent study suggests that endogenous E2 triggers macro- but its effect on M1 markers was very modest, and on the phage proliferation in ovariectomized mice [29]. Neverthe- contrary to other SERM7 and raloxifene, it did not affect less, we did not find E2 to modulate M1 type macrophage LPS-induced NFκB activity. LPS induction here is model proliferation in vitro. for bacterial endotoxin-activated inflammatory response Besides E2, SERM drugs such as tamoxifen, fulvestrant, [49], and therefore, it is tempting to speculate that SERM2 toremifene, and raloxifene have been previously linked with might suppress autoimmune reactions but still not the inhibition of tissue macrophage-mediated inflammation disrupting the innate immunity against pathogens. and monocyte differentiation and protection against neural Intriguingly, raloxifene stimulated LPS-induced but injuries [50, 53–57]. Our results support the idea that although inhibited TNFα-mediated NFκB activation. Raloxifene some SERMs, such as raloxifene and tamoxifen, may also treatment had also different trend between protein and promote an anti-inflammatory response, the novel com- mRNA expression of M2 surface markers in LPS- pounds SERM2 and SERM7 are more active in myeloid activated macrophages. Nonetheless, it has been previous- cells. However, the actual agonist/antagonist role of these ly presented that, on top estrogenic effects, raloxifene also SERMs on ERs and GPER in immune cells cannot be mediates ER-independent reactions in CD14+ cells and determined on the basis of our results. Previous studies lack may interfere their differentiation [50, 51] thus possibly unanimous results about contribution of various ERs on enabling apparently ambivalent changes in cell phenotype. macrophage polarization, suggesting that activation of ERα We will warrant the future in vivo studies to investigate or ERβ or even the membrane-associated GPER might alle- which SERM is the most efficient in polarizing macro- viate macrophage-related inflammation [11, 58, 59]. We pro- phages and inducing beneficial outcome in autoimmune pose that ERα signaling may have the most crucial role in M2 diseases. polarization. We base our claim on the facts that SERM2, It is noteworthy that SERM2 induced M2 phenotype SERM7, and raloxifene all had significant binding affinity to more effectively than E2, considering that estrogenic hor- ERα studied concentrations and ERα gene expression was mones in general are described in various studies mostly as the highest of ER subtypes in both resting and IFNγ/LPS- anti-inflammatory compounds, reviewed, e.g., by Straub activated monocytes, over 100-fold higher to ERβ.THP-1 [7] and Villa et al. [52]. To date, there has been several Lucia cells expressed somewhat different ER profile to pri- questions concerning the mechanisms of the anti- mary CD14+ cells, Gper1 being the most expressed ER. This inflammatory action of estrogen hormones that still remain difference is likely due to leukemic origin of THP-1 line, and unanswered [7, 52]. In addition to the general Banti-inflam- we cannot exclude that NFκB reporter activity assays here matory tone^ in which pro-inflammatory signaling is might be also affected by abundant GPER signaling. None- downregulated, E2-induced ER activation has been recent- theless, the central role of ERα in myeloid cell polarization ly associated with polarization of murine macrophages [16, and differentiation is suggested by other studies although the 25, 26]. However, the results reported in the literature lack of immune cell-specific methods somewhat hampers the suggest a diverse, context-specific role for E2 in macro- results [16, 27, 60]. phage activation. E2 promoted the M1 phenotype in rat There is an urgent need for new therapies for autoim- alveolar NR8383 macrophages and aggravated joint in- mune diseases as their global prevalence is approximately flammation [25], but on the other hand, ERα was required 5% in the total population thus affecting far over 100 for alternative macrophage activation, and E2 treatment of million people worldwide [61]. Novel potent drug mole- primary monocytes made them more prone to adopt M2 cules, especially cost-effective and well-tolerated, are high- phenotype [16]. E2 also accelerated inflammatory response ly required. Current SERM drugs lack several of the ad- in RAW264.7 cells by reducing the M1 > M2 transition verse effects of steroid hormones, and SERMs have been time after LPS shock [26] and interfered LPS activation by already used for decades as preventive and therapeutic impairing NFκBactivation[17]. Our results are in line agents thus providing plausible experience for prolonged SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells usage. We present here three SERMs exerting anti- cell compartment. Journal of Immunology 173: 2227–2230. https:// inflammatory effects in the myeloid cell microenvironment doi.org/10.4049/jimmunol.173.4.2227. 3. Saijo, K., J.G. Collier, A.C. Li, et al. 2011. An ADIOL-ERb-CtBP thus having a potential as novel pharmaceuticals to allevi- transrepression pathway negatively regulates microglia-mediated ate effects of chronic inflammation or autoimmune dis- inflammation. Cell 145: 584–595. https://doi.org/10.1016/ eases. Further understanding the impact of SERMs on j.cell.2011.03.050. 4. Lugo, L., A. Villalvilla, R. Largo, et al. 2014. Selective estrogen myeloid cell biology may also provide new insights into receptor modulators (SERMs): new alternatives for osteoarthritis? the benefits and disadvantages of ER-modulating drugs Maturitas 77: 380– 384. https://doi.org/10.1016/ currently used for, e.g., estrogen-sensitive cancers. Our j.maturitas.2014.01.016. results are still preliminary, but we intend to validate these 5. Paterni, I., C. Granchi, J.A. Katzenellenbogen, and F. Minutolo. 2014. Estrogen receptors alpha (ERa) and beta (ERb): subtype- observations in appropriate animal studies. Our aim is to selective ligands and clinical potential. Steroids 90: 13–29. https:// discover the most potent SERM to alleviate macrophage doi.org/10.1016/j.steroids.2014.06.012. dysregulation in chronic inflammation and autoimmune 6. Kovats, S. 2015. Estrogen receptors regulate innate immune cells – diseases. and signaling pathways. Cellular Immunology 294: 63 69. https:// doi.org/10.1016/j.cellimm.2015.01.018. 7. Straub, R.H. 2007. The complex role of estrogens in inflammation. ACKNOWLEDGEMENTS Endocrine Reviews 28: 521–574. https://doi.org/10.1210/er.2007- 0001. We thank Professor Kalervo Väänänen for helpful 8. Thomas, C., and J.-Å. Gustafsson. 2011. The different roles of ER subtypes in cancer biology and therapy. Nature Reviews. Cancer 11: discussions and advices and Mira Hannolin, Ketlin Adel, 597–608. https://doi.org/10.1038/nrc3093. and Hanna Haukkala for their technical assistance in 9. Polanczyk, M., A. Zamora, S. Subramanian, et al. 2003. The pro- performing the experiments. tective effect of 17beta-estradiol on experimental autoimmune encephalomyelitis is mediated through estrogen receptor-alpha. The American Journal of Pathology 163: 1599–1605. https://doi.org/ Funding Information This study was financially supported by The 10.1016/S0002-9440(10)63516-X. Finnish Funding Agency for Technology and Innovation (TEKES), 10. Dimayuga, F.O., J.L. Reed, G.A. Carnero, et al. 2005. Estrogen and decision diary no. 473/31/2015. brain inflammation: effects on microglial expression of MHC, costimulatory molecules and cytokines. Journal of COMPLIANCE WITH ETHICAL STANDARDS Neuroimmunology 161: 123–136. https://doi.org/10.1016/ j.jneuroim.2004.12.016. 11. Wang, C., B. Dehghani, Y. Li, et al. 2009. Membrane estrogen Conflict of Interest. Lauri Kangas and Tero Linnanen receptor regulates experimental autoimmune encephalomyelitis were employees of Forendo Pharma Ltd. at the time of through up-regulation of programmed death 1. JournalofImmunol- the studies and during the manuscript development. ogy 182: 3294–3303. https://doi.org/10.4049/jimmunol.0803205. 12. Relloso, M., L. Aragoneses-Fenoll, S. Lasarte, et al. 2012. Estradiol impairs the Th17 immune response against Candida albicans . Jour- nal of Leukocyte Biology 91: 159–165. https://doi.org/10.1189/ jlb.1110645. Open Access This article is distributed under the terms of 13. Haghmorad, D., A.A. Amini, M.B. Mahmoudi, et al. 2014. Preg- the Creative Commons Attribution 4.0 International Li- nancy level of estrogen attenuates experimental autoimmune en- cense (http://creativecommons.org/licenses/by/4.0/), cephalomyelitis in both ovariectomized and pregnant C57BL/6 mice through expansion of Treg and Th2 cells. Journal of which permits unrestricted use, distribution, and reproduc- Neuroimmunology 277: 85–95. https://doi.org/10.1016/ tion in any medium, provided you give appropriate credit j.jneuroim.2014.10.004. to the original author(s) and the source, provide a link to 14. Soldan, S.S., A.I. Alvarez Retuerto, N.L. Sicotte, and R.R. Voskuhl. the Creative Commons license, and indicate if changes 2003. Immune modulation in multiple sclerosis patients treated with the pregnancy hormone estriol. JournalofImmunology171: 6267– were made. 6274. https://doi.org/10.4049/jimmunol.171.11.6267. 15. Tai, P., J. Wang, H. Jin, et al. 2007. Induction of regulatory T cells by physiological level estrogen. Journal of Cellular Physiology 214: 456–464. https://doi.org/10.1002/JCP. REFERENCES 16. Campbell, L., E. Emmerson, H. Williams, et al. 2014. Estrogen receptor-alpha promotes alternative macrophage activation during cutaneous repair. The Journal of Investigative Dermatology 134: 2447–2457. https://doi.org/10.1038/jid.2014.175. 1. Vegeto, E., C. Bonincontro, G. Pollio, et al. 2001. Estrogen prevents 17. Toniolo, A., G. Paolo Fadini, S. Tedesco, et al. 2014. Alternative the lipopolysaccharide-induced inflammatory response in microglia. activation of human macrophages is rescued by estrogen treatment J Neurosci 21: 1809–1818. in vitro and impaired by menopausal status. The Journal of Clinical 2. Polanczyk, M.J., B.D. Carson, S. Subramanian, et al. 2004. Cutting Endocrinology and Metabolism 100: jc20142751. https://doi.org/ edge: estrogen drives expansion of the CD4+CD25+ regulatory T 10.1210/jc.2014-2751. Polari, Wiklund, Sousa, Kangas, Linnanen, Härkönen, and Määttä

18. Wegmann, T.G., H. Lin, L. Guilbert, and T.R. Mosmann. 1993. The Journal of Steroid Biochemistry and Molecular Biology 50: Bidirectional cytokine interactions in the maternal-fetal relationship: 313–318. is successful pregnancy a TH2 phenomenon? Immunology Today 34. Barsalou, A., W. Gao, S.I. Anghel, et al. 1998. Estrogen response 14: 353–356. https://doi.org/10.1016/0167-5699(93)90235-D. elements can mediate agonist activity of anti-estrogens in human 19. Simpson, E., G. Rubin, C. Clyne, et al. 2000. The role of local endometrial Ishikawa cells. The Journal of Biological Chemistry estrogen biosynthesis in males and females. Trends in Endocrinolo- 273: 17138–17146. gy and Metabolism 11: 184–188. https://doi.org/10.1016/S1043- 35. Kallio, A., A. Zheng, J. Dahllund, et al. 2005. Role of mitochondria 2760(00)00254-X. in tamoxifen-induced rapid death of MCF-7 breast cancer cells. 20. Voskuhl, R.R., H. Wang, T.C.J. Wu, et al. 2015. Estriol combined Apoptosis 10: 1395–1410. https://doi.org/10.1007/s10495-005- with glatiramer acetate for women with relapsing-remitting multiple 2137-z. sclerosis: a randomised, placebo-controlled, phase 2 trial. Lancet 36. Sousa, S., R. Brion, M. Lintunen, et al. 2015. Human breast cancer Neurology 15: 35–46. https://doi.org/10.1016/S1474- cells educate macrophages toward the M2 activation status. Breast 4422(15)00322-1. Cancer Research 17: 101. https://doi.org/10.1186/s13058-015- 21. Confavreux, C., M. Hutchinson, M.M. Hours, et al. 1998. Rate of 0621-0. pregnancy-related relapse in multiple sclerosis. Pregnancy in multi- 37. Qu, Q., H. Zheng, J. Dahllund, et al. 2000. Selective estrogenic ple sclerosis group. The New England Journal of Medicine 339: effects of a novel cholesterol level, and reproductive tissues in intact 285–291. https://doi.org/10.1056/NEJM199807303390501. and Ovariectomized rats *. Endocrinology 141: 809–819. 22. Jordan, V.C. 2003. Tamoxifen: a most unlikely pioneering medicine. 38. Barkhem, T., B. Carlsson, Y. Nilsson, et al. 1998. Differential Nature Reviews. Drug Discovery 2: 205–213. https://doi.org/ response of estrogen receptor α and estrogen receptor β to partial 10.1038/nrd1031. estrogen agonists/antagonists. Molecular Pharmacology 54: 105– 23. Goerdt, S., and C.E. Orfanos. 1999. Other functions, other genes: 112. alternative activation of antigen-presenting cells. Immunity 10: 137– 39. OECD series on testing and assessment no. 21 detailed review paper 142. (2002) Appraisal of test methods EENV/JM/MONO. 24. Ginhoux, F., J.L. Schultze, P.J. Murray, et al. 2015. New insights into 40. Carmeci, C., D.A. Thompson, H.Z. Ring, et al. 1997. Identification the multidimensional concept of macrophage ontogeny, activation of a gene (GPR30) with homology to the G-protein-coupled receptor and function. Nature Immunology 17: 34–40. https://doi.org/ superfamily associated with estrogen receptor expression in breast 10.1038/ni.3324. Cancer. Genomics 45: 607–617. 25. Kou, X.-X., C.-S. Li, D.-Q. He, et al. 2015. Estradiol promotes M1- 41. Plante, B.J., B.A. Lessey, R.N. Taylor, et al. 2012. G protein-coupled like macrophage activation through cadherin-11 to aggravate tem- estrogen receptor (GPER) expression in normal and abnormal en- poromandibular joint inflammation in rats. Journal of Immunology dometrium. Reproductive Sciences 19: 684–693. https://doi.org/ 194: 2810–2818. https://doi.org/10.4049/jimmunol.1303188. 10.1177/1933719111431000. 26. Villa, A., N. Rizzi, E. Vegeto, et al. 2015. Estrogen accelerates the 42. Mantovani, A., S. Sozzani, M. Locati, et al. 2002. Macrophage resolution of inflammation in macrophagic cells. Scientific Reports polarization: Tumor-associated macrophages as a paradigm for po- 5: 1–14. https://doi.org/10.1038/srep15224. larized M2 mononuclear phagocytes. Trends in Immunology 23: 27. Douin-Echinard, V., S. Laffont, C. Seillet, et al. 2008. Estrogen 549–555. https://doi.org/10.1016/S1471-4906(02)02302-5. receptor α,butnotβ, is required for optimal dendritic cell differen- 43. Wright, S.D., R.A. Ramos, P.S. Tobias, et al. 1990. CD 14, a receptor tiation and CD40-induced cytokine production. Journal of Immu- for complexes of lipopolysaccharide (LPS) and LPS binding protein. nology 180: 3661– 3669. https://doi.org/10.4049/ Science 249 (80): 1431–1434. jimmunol.180.6.3661. 44. Herz, J., A.J. Filiano, A. Smith, et al. 2017. Myeloid cells in the 28. Pelekanou, V., M. Kampa, F. Kiagiadaki, et al. 2015. Estrogen anti- central nervous system. Immunity 46: 943–956. https://doi.org/ inflammatory activity on human monocytes is mediated through 10.1016/j.immuni.2017.06.007. cross-talk between estrogen receptor ER 36 and GPR30/GPER1. 45. Heneka, M.T., M.J. Carson, J.E. Khoury, et al. 2015. Neuroinflam- Journal of Leukocyte Biology 99: 333–347. https://doi.org/10.1189/ mation in Alzheimer’s disease. Lancet Neurology 14: 388–405. jlb.3A0914-430RR. https://doi.org/10.1016/S1474-4422(15)70016-5. 29. Pepe, G., D. Braga, T.A. Renzi, et al. 2017. Self-renewal and 46. Roberts, C.A., A.K. Dickinson, and L.S. Taams. 2015. The interplay phenotypic conversion are the main physiological responses of between monocytes/macrophages and CD4+ T cell subsets in rheu- macrophages to the endogenous estrogen surge. Scientific Reports matoid arthritis. Frontiers in Immunology 6: 1–19. https://doi.org/ 7: 1–14. https://doi.org/10.1038/srep44270. 10.3389/fimmu.2015.00571. 30. Kramer, P.R., V. Winger, and S.F. Kramer. 2007. 17b-Estradiol 47. Röszer, T. 2015. Understanding the mysterious M2 macrophage utilizes the estrogen receptor to regulate CD16 expression in mono- through activation markers and effector mechanisms. Mediators cytes. Molecular and Cellular Endocrinology 279: 16–25. https:// Inflamm 2015: id816460. doi.org/10.1016/j.mce.2007.08.014. 48. Schultze JL, Schmidt S V. (2016) Molecular features of macrophage 31. Rettew, J.A., H.M.I. Samuel, and I. Marriott. 2010. GPR30/GPER-1 activation. Seminars in Immunology doi: https://doi.org/10.1016/ mediates rapid decreases in TLR4 expression on murine macro- j.smim.2016.03.009. phages. Molecular and Cellular Endocrinology 328: 87–92. 49. van der Poll, T., F.L. van de Veerdonk, B.P. Scicluna, and M.G. https://doi.org/10.1016/j.mce.2010.07.017. Netea. 2017. The immunopathology of sepsis and potential thera- 32. Messinger, J., B. Husen, P. Koskimies, et al. 2009. Estrone C15 peutic targets. Nature Reviews. Immunology 17: 407–420. https:// derivatives—a new class of 17B-hydroxysteroid dehydrogenase doi.org/10.1038/nri.2017.36. type 1 inhibitors. Molecular and Cellular Endocrinology 301: 50. Nalbandian, G., V. Paharkova-Vatchkova, A. Mao, et al. 2005. The 216–224. https://doi.org/10.1016/j.mce.2008.10.022. selective estrogen receptor modulators, tamoxifen and raloxifene, 33. Salomonsson, M., B. Carlsson, and J. Häggblad. 1994. Equilibrium impair dendritic cell differentiation and activation. Journal of Im- hormone binding to human estrogen receptors in highly diluted cell munology 175: 2666–2675. https://doi.org/10.4049/ extracts is non-cooperative and has a Kd of approximately 10 pM. jimmunol.175.4.2666. SERMs Promote Anti-Inflammatory Signaling and Phenotype of CD14+ Cells

51. Lee, S.A., E.Y. Kim, W.K. Jeon, et al. 2011. The inhibitory effect of cord injury in rats. Neurological Sciences 35: 1763–1768. https:// raloxifene on lipopolysaccharide-induced nitric oxide production in doi.org/10.1007/s10072-014-1828-z. RAW264.7 cells is mediated through a ROS/p38 MAPK/CREB 57. Poirier, A.-A., M. Côté, M. Bourque, et al. 2016. Neuroprotective pathway to the up-regulation of heme oxygenase-1 independent of and immunomodulatory effects of raloxifene in the myenteric plexus estrogen receptor. Biochimie 93: 168–174. https://doi.org/10.1016/ of a mouse model of Parkinson’s disease. Neurobiology of Aging 48: j.biochi.2010.09.007. 61–71. https://doi.org/10.1016/j.neurobiolaging.2016.08.004. 52. Villa A, Vegeto E, Poletti A, Maggi A (2016) Estrogens, neuroin- 58. Tiwari-Woodruff, S., and R.R. Voskuhl. 2009. Neuroprotective and flammation and neurodegeneration. Endocrine Reviews anti-inflammatory effects of estrogen receptor ligand treatment in er20161007. https://doi.org/10.1210/er.2016-1007. mice. Journal of the Neurological Sciences 286: 81–85. https:// 53. Komi, J., and O. Lassila. 2000. Nonsteroidal anti-estrogens inhibit doi.org/10.1016/j.jns.2009.04.023. the functional differentiation of human monocyte-derived dendritic 59. Spence, R.D., A.J. Wisdom, Y. Cao, et al. 2013. Estrogen mediates cells. Blood 95: 2875–2882. neuroprotection and anti-inflammatory effects during EAE through 54. Siracusa, M.C., M.G. Overstreet, F. Housseau, et al. 2008. 17beta- ERα signaling on astrocytes but not through ERβ signaling on estradiol alters the activity of conventional and IFN-producing killer astrocytes or neurons. The Journal of Neuroscience 33: 10924– dendritic cells. Journal of Immunology 180: 1423–1431. https:// 10933. https://doi.org/10.1523/JNEUROSCI.0886-13.2013. doi.org/10.4049/jimmunol.180.3.1423. 60. Elloso, M.M., K. Phiel, R.A. Henderson, et al. 2005. Suppression of 55. Barreto, G.E., M. Santos-Galindo, and L.M. Garcia-Segura. 2014. experimental autoimmune encephalomyelitis using estrogen Selective estrogen receptor modulators regulate reactive microglia receptor-selective ligands. The Journal of Endocrinology 185: after penetrating brain injury. Frontiers in Aging Neuroscience 6: 243–252. https://doi.org/10.1677/joe.1.06063. 132. https://doi.org/10.3389/fnagi.2014.00132. 61. Hayter, S.M., and M.C. Cook. 2012. Updated assessment of the preva- 56. Wei, H., and X. Ma. 2014. Tamoxifen reduces infiltration of inflam- lence, spectrum and case definition of autoimmune disease. Autoimmu- matory cells, apoptosis and inhibits IKK/NF-kB pathway after spinal nity Reviews 11: 754–765. https://doi.org/10.1016/j.autrev.2012.02.001.